Environmental sensor measurements and solar-voltage forecasts from a low-cost IoT device: Ulm and Zapopan field deployments

This dataset contains the raw environmental sensor measurements, on-device model predictions, firmware, and companion mobile application for a low-cost ESP32-based IoT device for environmental monitoring and 24-hour solar-voltage forecasting. It supports the article published in Electronics (2026). Two field deployments are included. A short hardware and firmware validation was carried out in Ulm, Germany. A 115-day campaign was conducted on the rooftop of Building J at the Universidad Autonoma de Guadalajara, Zapopan, Mexico, comprising 84 days of training (6 February to 30 April 2025) and 31 days of autonomous validation (May 2025). All sensor variables (temperature, relative humidity, luminosity as a 0-100 relative index, and solar panel voltage) were recorded at 15-minute intervals. The Zapopan files include the training measurements, the validation-phase sensor inputs, and the corresponding measured-versus-predicted voltage from the embedded feedforward network. The device firmware (.ino) and the Android application (.apk) are also included. See README.txt for a full description of each file.